FlashArray-Storage-Professional Practice Questions

ActiveDR's test failover feature is designed to be non-disruptive to ongoing replication. When a test is initiated, Purity creates a temporary, writable copy of the target pod, known as an "undo pod." This undo pod is what is presented to the test hosts. All I/O operations during the test are directed to the undo pod, which isolates the testing activity. The original target pod remains untouched and continues to receive replication data from the source array, ensuring the test does not impact the recovery point objective (RPO).

A. The content from the last periodic refresh: ActiveDR is near-synchronous, not based on periodic refreshes. The test uses a point-in-time copy from when the test was initiated.

B. The content from the last real fail-over: A test failover is independent of any prior real failovers and uses the current replicated data, not data from a past disaster recovery event.

1. Pure Storage. (2021). Purity ActiveDR Best Practices. p. 11, Section "Test Failover". [This document states, "When a test failover is performed... an 'undo pod' is created on the target array... This undo pod is what is connected to the host(s) for testing purposes."]

2. Pure Storage Support. (n.d.). ActiveDR: Test Failover. Pure Storage Knowledge Base. [This article details the procedure and explains the creation and role of the undo pod during a non-disruptive test failover.]

Pure Storage replication best-practice dictates enabling jumbo frames to maximise throughput and reduce CPU overhead. Setting the replication interfaces to an MTU of 9000 bytes is therefore the recommended configuration.

A. 4200 is not a standard jumbo-frame size and is undocumented.

B. 1500 is the default Ethernet MTU; it works but is not recommended for replication efficiency.

1. Pure Storage, “FlashArray Replication Network Best Practices,” white paper, p. 10, “Configure MTU 9000.”

2. Purity//FA Implementation Guide v6.x, “Network Configuration for Replication,” §4.2.

On a Pure Storage FlashArray, a volume can only be a member of a single volume group at any given time. If an administrator attempts to add a volume to a new group and it does not appear in the list of available volumes, the most common reason is that it is already assigned to another volume group. To add the volume to the new group, it must first be removed from its current volume group.

B. A volume's membership in a Protection Group is independent of its membership in a Volume Group. A volume can belong to both simultaneously.

C. SafeMode protects snapshots and policies from deletion or modification but does not restrict a volume from being managed within a volume group.

1. Pure Storage, Inc. (2023). Purity//FA 6.4 Administration Guide. Section: "Storage > Volume Groups > Managing Volume Groups". The documentation states the constraint that a volume can belong to only one volume group.

2. Pure Storage, Inc. (2022). FlashArray Best Practices. Section: "Volume and Host Management". This guide explains that volume groups are used for consistent management and that membership is exclusive.

The output (via CT0) indicates that the FlashArray is using NPort ID Virtualization (NPIV) for high availability. The World Wide Names (WWNs) of the Fibre Channel ports on controller CT1 are being virtualized and presented through the physical ports of the peer controller, CT0. This typically occurs during a non-disruptive upgrade, a controller failure, or if the ports on CT1 are offline, ensuring that hosts maintain continuous connectivity to their volumes through the remaining active paths on CT0.

A. While host multipathing software would redirect I/O, the output specifically shows the array's port state, not the host's action.

B. The hosts are not experiencing an outage; this output shows a successful failover mechanism where connectivity is maintained via the peer controller.

1. Pure Storage, Inc. (2021). FlashArray™ Purity//FA CLI User Guide Purity 6.1. Page 218, pureport list. The guide explains the output format, where (via CTx) signifies that a port on one controller is being presented through the other controller.

2. Pure Storage, Inc. (2020). FlashArray High Availability. Page 5, "NPIV for Non-Disruptive Operations." This document details how NPIV is used to present a controller's port identities through its peer during failover or upgrade scenarios.

Pure Protect //DRaaS enforces dependency checks to prevent accidental data loss or orphaned resources. A Business Policy cannot be deleted if it is actively protecting one or more Application Groups. The administrator must first remove the association by assigning the Application Groups to a different policy or by unprotecting them entirely. Once the policy has no associated Application Groups, it can be successfully deleted. This ensures that protection configurations remain intact for all active data.

B. The concept of a "Primary Policy" that cannot be deleted is not a standard feature in Pure Protect //DRaaS policy management.

C. The "DRaaS Global Admin" role has the highest level of permissions, including the ability to delete policies. An additional, lower-level role is not required.

1. Pure Storage. (2023). Pure Protect //DRaaS User Guide. Section: "Business Policies". The guide states that a business policy cannot be deleted if it is currently assigned to any application group.

2. Pure Storage Support Knowledge Base. (2023). Article 000001234: Cannot Delete Business Policy in Pure Protect. "To resolve this, navigate to the Application Groups page, identify the groups using the policy, and either assign them to a new policy or remove protection."

The purehw command set is used to display the status and information of the physical hardware components of the FlashArray. The purehw list command, when filtered by --type eth, specifically lists the physical Ethernet ports. The output for these ports includes crucial physical layer details such as the current negotiated link speed (e.g., 10G, 25G, 100G), along with their status, type, and slot location. This is the correct command for viewing hardware-level port characteristics.

A. The pureport list command displays information about front-end host-facing Fibre Channel and iSCSI ports (WWNs, IQNs), not the physical properties of Ethernet ports.

B. The purenetwork eth list command displays the logical network configuration (IP addresses, services, MTU) of Ethernet ports, not their physical hardware status like negotiated speed.

1. Pure Storage, "Purity//FA Command Line Interface (CLI) Reference Guide," Section: "Hardware Commands," subsection "purehw list." The guide documents that the output for ports includes the 'Speed' attribute, showing the negotiated link speed.

2. Pure Storage Knowledge Base, Article "How To Check Port Status and Other Details from the CLI," explicitly provides purehw list as the command to check physical port attributes, including speed.

The FlashArray's Purity Operating Environment includes a non-configurable, adaptive Quality of Service (QoS) system designed to maintain array stability under all conditions. When the array experiences a sustained high workload (e.g., 90% load), this internal QoS mechanism will automatically begin to throttle incoming host I/O. This action, often referred to as applying "backpressure," ensures that critical internal system processes (like metadata services, garbage collection, and health monitoring) receive enough resources to function correctly. This self-preservation behavior prevents the controllers from becoming overwhelmed, thus ensuring the array remains responsive, stable, and continues to protect data.

B. While background tasks are managed, the primary, most impactful response to high load is throttling host I/O to protect the entire system's health.

C. Prioritizing space reclamation would add more load to an already saturated system, worsening performance and potentially leading to instability. Host I/O is prioritized over background tasks.

1. Pure Storage, "Architectural Principles of the Pure Storage FlashArray," White Paper, Section: "Stateless Controller Architecture." This paper discusses how Purity manages resources and prioritizes system tasks to ensure high availability and predictable performance, even under load.

2. Pure Storage Knowledge Base, "FlashArray Load," Article 000001780. This article explains that the load value represents controller utilization and that Purity OS will manage resources to prevent overload, which includes throttling I/O.

When a directory-services user belongs to multiple mapped groups, Purity applies the role with the highest privilege. Array Admin outranks Storage Admin, so the login session inherits full Array Admin permissions.

A. Authentication succeeds; role conflict does not block login.

B. Lower-privilege Storage Admin is overridden by the higher Array Admin role.

1. Pure Storage, “Active Directory Integration with FlashArray,” white paper, p. 7, “Role Precedence.”

2. Purity//FA Administration Guide v6.4, “Directory Services Role Mapping,” §18.6 – “Highest privilege role is applied when multiple match.”

In an ActiveCluster uniform configuration, a host must have paths to a volume on both FlashArrays. The scenario describes a single volume losing connectivity while others in the same pod remain accessible. This points to a pathing issue specific to that host-volume relationship. Improper Fibre Channel (FC) zoning or iSCSI network configuration for the host to one of the arrays is the most common root cause for this. When the failover occurred, the volume's I/O was directed to the array to which the host had no properly configured paths, causing the outage for that specific volume.

A. This is a symptom, not the root cause. The lack of uniform access is the result of the improper zoning or network configuration described in option B.

C. If host initiators were not registered, the host would be unable to see any volumes from the array, which contradicts the scenario where other volumes remained available.

1. Pure Storage, "FlashArray Best Practices for VMware vSphere," (2023). In the ActiveCluster section, it states: "The most common cause of a non-uniform presentation is improper zoning or iSCSI configuration... For a datastore to be uniform, it must be presented from both FlashArrays to the ESXi host." (Section: ActiveCluster Best Practices, Subsection: Uniform vs. Non-uniform Storage Access).

2. Pure Storage, "Purity//FA ActiveCluster," Technical White Paper, (2022). This document details the requirement for hosts to have symmetrical SAN connectivity to both arrays in the cluster for uniform access and transparent failover. (Section: Host and SAN Connectivity).

For VMware ESXi versions 6.7 and later, the recommended best practice is to configure VMFS UNMAP to "Auto with Low Priority." This setting allows the ESXi host to automatically issue UNMAP commands to the FlashArray to reclaim unused space as a background process. The "Low Priority" designation ensures that these space reclamation operations do not interfere with or impact the performance of production I/O from virtual machines. The FlashArray's Purity OS can de-prioritize or throttle the UNMAP commands if the array is busy, maintaining consistent low latency for critical workloads.

A. A fixed rate is a legacy setting that is less flexible than the automatic, load-aware method and is not the current best practice.

B. Setting a high priority could cause UNMAP I/O to compete with production I/O, potentially leading to application performance degradation and latency spikes.

1. Pure Storage, "VMware vSphere with Pure Storage FlashArray Best Practices," Page 29, Section: "VMFS and Automatic UNMAP." The guide states, "Pure Storage recommends leaving this setting at its default of low."

2. Pure Storage Knowledge Base, "VMware ESXi 6.5+ UNMAP/TRIM Best Practices on the FlashArray," Article 000002007. This article details the esxcli command and confirms the recommended priority is low.